How practical is this Feature Request? I’d like to discuss before proposing.
Let’s just let the user enter parameters for an array of crosshairs and burn them, have them manually click on them, and use that for a matrix.
It seems like it’s just basic interpolate within the triangles (or squares) and get known, fixed, pixel-accurate points and as high an interpolation accuracy as the user wishes if they want to click through more calibration points all day.
24 or 48 points across the bed wouldn’t be that much time spent, really. Offset seems to make more sense than square pattern.
I’d like to avoid the automated registration point burning. The rasters take longer than they need and aren’t better for accuracy. When doing my later testing I just burned 15mmx15mm lines on paper, a bit out of focus to widen the burn so they appear bolder on the camera. More than 10x faster. Numbering seems unnecessary, as they’re spaced out enough that the software will know which one they are by basic location.
On a practical front, I’m thinking we need to be able to burn the crosshairs and align them in groups. That is, I’ve got a 1.6m x 1.0m bed. Even if I use paper, that’s a silly number of sheets. It’d be nice to lay down one or two sheets and burn 6 crosshairs, click them, then move the sheet over and offset a bit when I place the paper so the new marks won’t be confused with the existing ones, burn another 6 on the same sheet, find and click the new ones, and just walk it across the bed.